The Modeling Core will provide a central resource for all computational modeling across projects. Modeling plays a critical role in this program of research because it links theoretical, behavioral and neuroimaging approaches to understanding the relationship between reading and speech. The triangle model of reading is central to this enterprise because it provides a mechanistic explanation of how mappings among spelling, phonology and meaning are learned and used to support reading. Prior work in this framework has provided insights into how phonological and semantic processing contribute to literacy outcomes. All four projects will make contact with the triangle model, integrating insights and new advances from models of phonology (Project II) and sentence processing (Project IV), extending the model to new languages (Project III) and examining how different patterns of constitutive deficits in a variety of language abilities can have different impacts across languages (Project I). This integrative approach will open up new directions in the theoretical and empirical work, and make progress toward a general theory relating the development and use of reading skill to speech and language. Finally, interaction with the Neuroimaging Core (Core B) will produce new ways of interpreting and analyzing fMRI data. The models are pitched at a level that is well-suited to this, because they provide proxies for both behavioral outcomes such as response time and accuracy, as well as internal states that can be directly related to brain activity. Thus, the Modeling Core will coordinate activities across projects and cores that are central to unifying the neurobiological and computational models of reading and its relation to speech